So, been looking at numbers in my DB.

For the most part I have been sticking to solved ranges. I have gtowizard as well as the red chip solved ranges.


When I follow those to a T, I'm only attempting to steal blinds about 30% and have a 60% success rate.



But, its always a topic of discussion on the interwebs about how much you should steal or defend.



So, at what point do you usually deviate? I.E. how many hands do you usually like to see on a player to know you can widen your steal attempt. Or if you're playing in something like a zoom pool, what stat/s are you looking to indicate you should widen your steals?


Thanks

I'm also folding to steals about 70%, which seems a bit high.

Stealing the blinds 30% is very low if we're only counting BTN and SB, not solver approved. Maybe from SB specifically if you have limps, but I recommend for now you simplify your game to not limping and just opening the same hands on BU and SB.

I would shoot for around 150 cases of facing a steal attempt on HUD, and for a +-5% deviation, you can start adding borderline hands and eventually once it gets to an extreme (like your defense), you can open any 2 cards

This is why I was asking. 30% seems very low to me. I'm use a 3bet or fold strategy only currently. Only calling from BB. None from SB. Only only opening or folding from SB in that situation.


Mostly using the red chip solved ranges since they are very easy to access and are 3bet or fold except from BB.


I'll keep doing what I'm doing and see where it ends up with more data.


Thanks

If yot steal success is really 60% you can open 100%.

Thanks. That was my next question. What is the threshold for say a zoom pool folding to steals that you can damn near open ATC on button?

I fluctuate around 50-60%. *But* that's on a 30% steal attempts. So, even 60% of that aint much.




Will keep eye on it. At what point would you feel comfortable with a sample size of hands to start making adjustments? 50k, 100k?

BvB If BB is folding like 50% or more you can open pretty much ATC if they dont 3b something crazy.
You can open wider from the start because only top regs defend enough and rest of the pool under 3b and folds too much. Esp in zoom pool where its so easy to just fold and go to the next hand.

In conjunction with fold to steal, I use VPIP/PFR for small samples (<200 hands) as an indicator of whether or not I can steal wider. Generally, if I see a high FTS (>75%) but it's only over a small sample, I use the VPIP/PFR to tell me whether or not I'm likely dealing with a nitty player who would be more inclined to fold.

I also look at 3B% as well, more marginal opens = more folds when facing 3b.

Look at your ranges on the preflop grid. The space between the weakest hands that youre including, and the strongest hands you're not including is assumed to be representative of the 0ev profit margin. I have the loose theory that if the opponents are allowing me to profitably add the next best available combo to my range, then it follows that I may add a combo to every row on the grid along the margin. Likewise, if an opponent is causing me to tighten up, I dont just tighten one row. I tighten the whole margin.

The question of "when to deviate" is easy, the answer is: ALWAYS!

Lots of players on these threads are in the micros. In the micros there is loads of chips available from steals. I’d literally steal every hand unless there an opponent who discouraged me. Lots of you guys will over think it: “If I steal every hand, they will notice and start playing back at me”. They won’t!! They might notice, but they will just get annoyed, they won‘t adjust from their trusted strategy. The micros is insanely easy. Most the regs are nits. So just steal, cbet once, and make easy money. It is literally so easy that it is boring.

Around 1 in 5 players in the micros are the complete opposite, and instead of folding loads, they call loads. It’s not hard to recognise these guys and then you just switch it up against them. I’d limp a lot against these guys, as your implied odds are excellent when you hit the flops hard.

As you move up to the low stakes (which you will do very quickly if you follow that very simple approach) you find that some players are playing attention to your plays. These guys will adjust to your steals by 3betting you loads, but what they wont do, is pay attention to how you are playing against them specifically. If you simply tighten up when they are in the blinds, you can rinse them when they 3bet against you. Again, these guys will just get angry with you, they wont actually know how to contend with your strategy.

Stealing is great, but as you move up to the mid-high stakes, the opponents will be extremely well versed in defending against steals. What you need to do against these guys is add a new dimension to your stealing strategy by adding LIMPS.

If you are limping 10% of the time, and stealing 50%, you’re able to switch up your value to bluff ratios without the opponent seeing any change to your strategy. You could limp your weakest hands if the opponent folds to limp-raises (which they all do). You can limp some of your weaker improving hands, like K5s, if the opponent is 3betting lots. If the opponent notices and starts playing back against our limps, and avoiding our opens, we just switch it up, we limp raise some of those nutsy hands, and as we are still only limping 10% it’s impossible for the opponent to notice until we reach showdown and he loses a stack. Even then, he will rack his brain trying to work out what the hell you are doing. They will probably settle on: “He must be limping a balanced range” and will go back to their solver to find out how to contend with it.

What they don’t realize is that we don’t actually intend to make much, if anything, from our limps, but instead, we’re trying to control the strength of our range for when we do steal. We are complicating things, so that their GTO strategy will have to be far more advanced for them to contend with us. These guys usually associate a limp-raise with the nuts, they usually associate a limp from the button with a fish, so, we are playing with their assumptions to mess with their heads and encourage them to make errors.

I came up with a genius strategy that I called “The Anti-GTO Strategy”. Obviously you can't actually beat GTO, but you can do something really weird that takes profit from the GTO players in a very obscure way that is difficult for them to contend with. The way the GTO players learn to play is by learning to follow the GTO strategy for certain situations, in those situations they believe that they are unbeatable, but that is NOT TRUE. Their overall strategy is unbeatable, but in any one situation they can be beaten. For example, if I only ever open AA on the button, when I open the button, I will absolutely obliterate the GTO players. If I take this extreme approach I will lose profit by not opening all the other hands in the GTO button open range, but, when I open, I will rinse them.

So, more advanced opponents are very good at contending with steals, but they are less well versed in dealing with limps. What we do then, is open/steal a tighter range than GTO to increase our profit when we open, and then instead of just folding those extra hands we increase our limp range to play against the GTO player in situations that they are not well versed. If we only tighten our open/stealing range slightly, we can make easy profit against the GTO regs when we open, then we fight it out to regain our lost profit in the limp pots.

Stealing is very simple in the lowest and easiest games, but even in the highest games there are complicated limping strategies that you can use to confuse the most well trained of opponents. Nobody else in the world understands the anti-GTO strategy. I designed it, and when I posted about it, some of the biggest names in poker told me I was an idiot. Which means, even when I explain it to them in plain english they don’t understand it. These poor guys don’t stand a chance of understanding what I am doing if all they see is me limping a little more than GTO.

So, I have almost 18k hands (working on 40k hands to start coaching from Aner0.


I'm sitting at 28 for my Att to steal. As far as I know, pt4 defines that as CO, BTN, and SB.
I'm sticking very, very close to the solved ranges I'm using. Is this somehow still too low?



I have a 45% success rate.



If the 28% isn't "solver" approved, any idea what's influencing this? Or just small sample and the preflop action ahead of me is dictating it smaller?

I think this is all only true for players who insist on playing an equilibrium strategy. Which is quickly becoming not the norm.



Without getting into semantics, the "optimal" strategy always takes player tendencies into account and adjusts.



Exploitive or "non gto" play is literally just a solver with node locking. If you give it the correct input for node locking, it will spit out the strategies you are discussing above.




Anyone not using node locking a very high amount is definitely not playing "optimal game theory." They are just playing an equilibrium strategy.






And your example about AA from BTN, while I understand is simple, only works against players who never adjust and don't node lock/exploit.



Its the exact same in a RPS game. If your opponents only play paper, almost no one would continue an equilibrium strategy. They would all start playing scissors. And the really good players would play scissors only like 80% of the time to attempt to not bring V's attention to their exploit and change strategy.

And most of the time, these types of players are on extreme opposites of the bell curve.

You have guys that will just be maniacs who are on one side. Zero theory involved.
Then you have players who attempt to play a pure equilibrium strategy which is on the very far opposite side of the bell curve


And thusly, adapting our main strategy to exploit either of those is not a good idea. You want to be using a general strategy that targets where the most players are. Which is 1-2 standard deviations from center.

Just look at RFI and Steal succ as function of position. Also state what your rfi sizings are for sb/btn/co

2.5bb from CO/BTN.

3bb from SB

Solvers dont tell you how far the really good players should adjust for their strategy to remain camouflaged. Node locking is a GTO strategy against an opponent who has some unrealistic disability, this isn't the only way to exploit, not by a long shot, it's only Level 2 for a start, which is very basic. To exploit you need to get into the players head, understand what the player is like, and then exploit their weaknesses. For example, my brother can't beat me because he's scared of big pots, there is no node lock that covers this sort of ailment. If you see a villain with 80% fold to steal, it's obviously a good idea to steal against this opponent, but that stat tells us way more about this opponent than the fact that they usually fold often to steals. And, stats showing 80% Fold to Steal doesn't mean his strategy against you is 80%, nor that it will stay at 80% either. Node locking is a good way to learn how to exploit safely against a one particular strategy, but it's not realistically a good way to exploit, it won't even maximize your profit from the exploit, and it's far more complicated than a natural exploit too, I mean, how many years would it take you to find a strategy on a solver that effectively exploits my brother? For me it's easy, I just make myself look scary, like I'm going for stacks, if I want him to fold anything but the nuts.

I don't think you understand what I was saying when I used the AA as an example. It is very hard to adjust against my Anti-GTO strategy. Node locking most certainly won't be enough, as the villain won't have a clue what strategy I'm using, and even if he somehow works it out, the strategy is fluid and always changing. A good exploitative player is always fluid, and always changing strategy depending on the opponent's weaknesses and what the opponent has seen. Solvers dont account for this. If you steal every button, it won't be long before anyone but the very worst players begin to adjust. Solvers can't predict any of that, because solvers can't empathize, the human mind has far more tricks up it's sleeve than any computer. We can touch on the infinitely complex nature of our opponents strategy, but computers can only follow finite strategies, they are still inadequate compared to us when it comes to exploiting.

Huh? That's what node locking is

Riiight... show me the node lock for "Scared of big pots"... I'm getting bored 2+2. This post is just foolish.

Nope. I've decided i cant be bothered to help you

Darn it!... I was hoping next you'd show us a solver that teaches us how to beat women!.. I apologize if any men out there find that offensive, I do want to beat your asses too... How about Gingers instead, we all hate losing to gingers, can't you show us a solver that teaches us how to beat people with this hair impairment? No? I guess I was right then, solvers cant teach you everything.

Im hesitant to engage since you are clearly such a troll but I want to tackle one point.


"how many years would it take you to find a strategy on a solver that effectively exploits my brother? For me it's easy, I just make myself look scary, like I'm going for stacks, if I want him to fold anything but the nuts."


If an opponent is folding everything but the nuts (or near nuts) on the river to a big bet, it would take <10min to node lock that range for him and calculate the optimal betting strategy against that range. Node locking is as accurate as your perception of their range. If you get their range correctly, its optimal, but if you get their range wrong it can result in losing plays.

Even without node locking, the solver is only as accurate as the preflop ranges given to it for each opponent. People tend to input preflop ranges that are also generated by solvers because those are the common ranges that players study and try to implement, but the solver strategy against a super tight preflop range is going to look much differently than a solver strategy against a super loose preflop range.


There isn't a single player who studies game theory that believes you shouldn't deviate at times to exploit players who make obvious mistakes. Whether that means not bluffing the calling station or overfolding vs someone who never bluffs.


A solver is simply a EV Maximization calculator and it is only as accurate as its inputs. The GTO strategy is the strategy that maximizes the EV of both players across the full game tree, it cannot be beaten by deviations from the opponent although it wont maximize EV against those deviations.

My brother won't always fold everything but the nuts. That isn't exactly what I said. He's a good player so wont fold every board to aggression, but; if he ever leaves the nuts from his perceived range I usually will go to defcon 7 and nuke the b******, if the board is very scary and he usually has a pair/2p type hand, I will often go kamikaze, regardless of my hand strength. You cant node lock any of that.

Solvers dont go for max EV either, not when node locked or otherwise. They still exploit and counter exploit until neither player can exploit any more. When they are node locked it just gives one player a bigger advantage, it finds the GTO solution in a toy game where the opponent is incapable of certain moves. This is not the most profitable way to exploit. It is a safe way to exploit against holes in a player strategy who tries to use GTO but is incapable of adjusting, which is kinda useful in this slightly common situation nowadays, but it is still not really practical. And you sir are evidence that this can cause lots of confusion, making it in many ways more difficult for players to understand how best to exploit in certain situations.

The best argument I've heard for using node locking on a solver is to use it as a reference point, so you can understand how a GTO strategy might change based on different holes in the opponents strategy. However, the confusion created far out weighs the benefits. For a start, it's not very important to learn how a GTO strategy might change based on different holes in the opponents strategy, it's probably much more important to learn the maximum exploitative strategy instead. And the confusion created is monumental. This is only Level 2 exploitation and so players who use solvers are rarely going to reach the all important level 3.

The best argument I personally came up with for using the node lock feature on solvers is to use them to see how an solver trained opponent might adjust to perceived holes in your own strategy. This is Level 3 and is an exercise I could get behind. Still though, the most important thing for any player to learn is the fundamentals of exploitative theory, as this enables them to use their own natural calculations to make the most money possible in any situation. This process has evolved for billions of years and is far better for us to use than a solver node lock is. When you realize that the solver node lock training actually hinders your ability to use this thought process you realize is it almost damning, as any benefits will be outweighed by the disadvantages.

To me it's all quite simple. Master exploiting, reach the mid stakes. Make at least $50 an hour from exploiting, which is rather easy. Then you are ready to look at GTO and solvers. Your understanding of theory will have advanced enough to negate the disadvantages and so there is only really benefit from learning to accurately understand the GTO strategy. I myself would do this if my career swayed back towards playing more Poker. However, I feel that I should protect my ability to comprehend more theory, and so I still stay miles away from GTO.

"Solvers dont go for max EV either, not when node locked or otherwise."


This statement is soo soo wrong yet written with such confidence.


A solver calculates a strategy that maximizes EV throughout the entire game tree, given certain inputs for each player. That is literally all it does. At default, it assumes that both players are playing optimally, but you can adjust ranges and tendencies to attempt to model any type of player. And if your model of your opponent's strategy is accurate, then the solver will calculate the max EV line against them.


"if he ever leaves the nuts from his perceived range I usually will go to defcon 7 and nuke the b******, if the board is very scary and he usually has a pair/2p type hand, I will often go kamikaze, regardless of my hand strength. You cant node lock any of that."

Sure you can. This is simply the case of your opponent having a capped range, and you having a nut advantage. Solvers apply max pressure to a capped range all of the time. If you remove the nuts from your opponents range and cap him at 2pair on a 4 straight/3 flush board, then your strategy is going to be super polarized with very large bets.

Ohhh a challenge from the Live Tourney Pro. Impressive!
Bravery is a great trait, provided you can bow down when you're beaten...
I may well be misunderstanding this myself, but let's see...

...

"Solvers dont go for max EV"

https://www.tournamentpokeredge.com/...oitative-play/

Are you sure solvers dont use the minimum exploitative strategy in that article? You know the difference between the two?

...

"If he doesn't have the nuts I will go to defcon 3..."

Essentially what I was saying here is that you can't feed characteristics into a solver, you can't input into a solver the range that the opponent thinks you hold, nor do solvers account for any potential changes to the opponents strategy.

- A players characteristics do lead to certain strategies being applied, and these strategies can be fed into a solver, but there is A LOT lost in this translation.

- Solvers haven't advanced to a point where you can input a different range for yourself containing the hands the opponent thinks you hold. This would lead to him using a different strategy, which we could then design our own strategy to exploit.

- Potential changes is huge too. The way I understand it, solvers exploit and counter exploit until neither player can make any improvement without the other player changing strategy, but in the real world there is only one play you can make, adjusting and counter adjusting happens over time and it is the speed we can make these adjustments that determines the EV we gain over time.

The perceived range is a Level Three factor that we consider during a hand, potential futures is a factor that is first introduced once a strategist becomes unconsciously capable of Level 3, we humans use our empathy to think at Level 3 and this empathy enables us to imprint on our opponents some of the characteristics that we recognize in ourselves which we can then use to help us gage the opponents strategy and how it might change in response to our actions. This is all Level Three stuff, and solvers are Level Two.

The tone of your post makes me think you are also a L2 player. If, at the tables, you don't think about your perceived range as much as the opponents range, you will be a L2 player. Before you respond, please spend some time thinking about your perceived range, or maybe even describe to us how it is different to your actual range?

I didn't say it maximizes EV for a single hand on a single runout. It maximizes EV for your entire range throughout the entire game tree.

This article isn't discussing solvers at all, just Exploitive Play in the abstract. You are clearly a proponent of maximally exploitive play, going balls to the wall trying to "outplay" your opponent. Here is a quote from that section explain my objections to this method

"Firstly, we never truly know our opponent’s strategy. We can make some very accurate estimates from time to time, but unless they actually tell us beforehand, we really don’t know for sure. Even if they do tell us, they might be bluffing! This means that in real-game situations, any exploitative adjustment to our play that we make is likely to be at least slightly incorrect. The trick is to make sure that we ground our adjustments in enough information that we’re right more often than we’re wrong.


Minimally exploitive play is how I recommend learning and implementing it. A solid baseline game based in theory with adjustments made on the fringes to increase EV from your opponent's obvious mistakes.


And yes I understand how your "levels" concept works (I actually read your pdf btw... could have used an editor). Its not some revolutionary idea, its been around a long time and before solvers were created and we started learning the correct way to play. I was a multitabling mid stakes cash game reg pre solvers, I understand how to exploit. I just dont think its the foundation to a solid poker game in 2022, its the seasoning that takes your game to the next level if built upon a solid base.

Level 1 - Your hand

Level 2 - Your opponents hand (range)

Level 3 - What your opponent thinks you have

Level 4 - what your opponent thinks you think they have

... forever until you get to GTO. You want to be exactly one level above your opponent at all times. If they only think about their own hand, you want to think about their range. If they are also thinking about your range, you want to think about what your range looks like to them. If you are more than one level ahead then you are thinking about things that have no impact and may end up "leveling yourself" and making a poor decision.


The difference between my "perceived" range and my actual range is entirely up to my opponent. If they perceive my range to be way different than it actually is, then they will make mistakes. But in my eyes, there is no difference, my range is my range. If my range contains a lot of flushes/straights, then I am going to play a board with flushes/straights aggressively, regardless of what my actual hand is. They are going to "perceive" me as having a lot of strong hands because I do. Now if you find an opponent who calls you down super weak despite this, you know that they are not perceiving your range accurately and can make some adjustments vs this players, such as removing some bluffs from your range so his calls lose even more money.